Embed Size (px)
Citation preview
1
Working Paper Series Paper No.1/2004-PC
PENSION LIABILITIES OF THE CENTRAL
GOVERNMENT: PROJECTIONS AND IMPLICATIONS
SIBANI SWAIN PRONAB SEN
PERSPECTIVE PLANNING DIVISION PLANNING COMMISSION GOVERNMENT OF INDIA
JULY, 2004
2
PENSION LIABILITIES OF THE CENTRAL GOVERNMENT: PROJECTIONS AND
IMPLICATIONS
Sibani Swain Pronab Sen*
ABSTRACT
The commonly held belief that the Central government pension bill has the
potential to reach an unsustainable level does not appear to be based on any realistic assessment of such liabilities in the future years. The future pension liabilities of the Central Government would be primarily driven by the structure and behaviour of employment in the Central Government in the past and the structure of prevailing pension system. This paper applies theoretically established models to the available time series on employment to estimate the growth in number of pensioners and make realistic projection of the future pension liabilities of the Central Government comprising five accounting departments such as Railways, Defence, Posts, Telecommunication and Civil. On the basis of these estimates it is concluded that if the present pension system prevails, the future pension liability of the Central Government would be reasonably sustainable.
* The authors are respectively Director and Adviser in the Perspective Planning Division of the Planning Commission, Government of India. The views expressed in this paper are the personal views of the authors and not necessarily those of the organization to which they belong.
3
Pension Liabilities of the Central Government: Projections and Implications
Sibani Swain Pronab Sen
Introduction
A consensus appears to have been reached among economists, financial experts, planners and policy makers that the present non-contributory and defined benefit nature of the pension system for all government employees is unsustainable and needs to be changed forthwith. Thus, pension reform has now become one of the key priorities in the Central government’s fiscal reform agenda. The concern relating to government pension has its origin in the alarming increase in the manpower cost of the government in recent years. At present, the salary and pension bill of the employees of the Central government and State governments put together constitute about 9.2 percent of the gross domestic product (GDP)1. As can be observed from the Graph-1 below, the salary and pension bill of government employees has increased from 7.9 percent of the GDP in the year 1990-91 to the present level, the rise being attributed mostly to the implementation of Fifth Central Pay Commission (FCPC) award from 1997-98 onwards. The consequent fiscal stress is obvious.
Graph 1 Salary and Pension of Centre and States as % of GDP
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
1990-91
1991-92
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99
1999-00
2000-01
2001-02
Salary Pension Total
1 This does not include the salary and wage-bill of the employees of the local bodies
4
However, it can also be observed that a gradual downturn has already started in the ratio of government expenditure on salary and pension to GDP. The downturn is more pronounced in case of the Central government. The position of the Central government, in this regard, appears to be fairly comfortable as shown in Graph 2. The salary and pension payment to the Central government employees as a percentage of GDP at present (2001-02) has come down to the level of 1990-91 after peaking in 1998-99. The wage bill, taken separately, is lower in the year 2001-02 than it was in the beginning of 1990s. This is partly attributable to the Central government’s effort to gradually prune down the number of employees and partly to the recent corporatisation of the telecommunication department, among others.2 The Central government liabilities on account of the payment of retirement benefit to its employees, though still high at about 0.9 percent of GDP, has also started declining.
Graph 2 Salary and Pension Bill of Central Government as % of GDP
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
1990-91
1991-92
1992-93
1993-94
1994-95
1995-96
1996-97
1997-98
1998-99
1999-00
2000-01
2001-02
Salary Pension Total
Thus, the commonly-held belief that the pension bill has the potential to reach an unsustainable level, or perhaps even exceed the wage bill, does not appear to be borne out by the recent trends or to be based on any realistic assessment of such liabilities in the future years. Nevertheless, there is near unanimity on the desirability of introducing pension reforms on grounds of both fiscal prudence and equity. 3 It appears that the high rates of growth recorded by
2 The corporatisation of the operating arm of the Department of Telecommunication has taken the salaries of these employees off the government’s rolls, but the pension liabilities continue to be borne by government under the pact entered with the unions. 3 The equity argument stems from the perception that the existing pension scheme for government servants is vastly more generous than anything available in the private sector or even in public sector enterprises.
5
pension payments since the mid-1980s continues to colour perceptions, without any significant effort at assessing the causes or evaluating the likelihood of such trends persisting in the future. This may be of no major consequence if a decision to reform the government pension system has already been taken. However, transition to a contributory pension scheme for government employees from the present non-contributory defined benefit pension system does not seem to have been an easy task. Even after more than two years since the budget announcement (Union Budget 2001-02) regarding implementation of pension reform, the basic principles of the new pension scheme are yet to be firmed up. More importantly, resistance from the unions is yet to be faced. Under the circumstances, therefore, there appears to be some utility in at least clearing up cobwebs surrounding the numbers. If the government for whatever reason is unable to implement a new pension scheme, it should at the very least have a reasonably accurate idea of what its future fiscal liabilities will be. A simple-minded projection of pension liabilities from the historical trend is obviously erroneous, since there is no reason to believe that government employment, which determines the potential stock of future pensioners, has followed a stable trajectory. Indeed, the evidence is to the contrary. By all indications, the maximum growth in government employment took place in the 1950s and 1960s, and slowed down significantly in the 1980s and 1990s. Moreover, the historical growth path observed in pension payments does not appear to correspond to any economic logic. Neither the rate of inflation nor the rate of increase in the number of pensioners (estimated) in the past supports the growth pattern of pension liabilities (more than 20% annually) that have been experienced during 1980s and 1990s.4 It appears essential, therefore, that a more rigorous methodology be used for making such projections.
In the recent past, some attempt has been made to make projections for the future pension liabilities of the Central government under a ‘no reform’ or ‘business-as-usual’ assumption. At least two official reports which make such projections for the Central government, comprising of five accounting departments, namely Civil, Defence, Postal, Railways and Telecommunications, are available. First, the report of the “Working Group on an Assessment of Government of India Pensionary Liability” submitted in 2001 (referred to as the Working Group in rest of the paper) projects an annual growth rate of pension of about four per cent for the period 2000-2010 for all accounting departments of the Central Government taken together. Second, the report of “High Level Expert Group on New Pension System” submitted in 2002 (henceforth referred to as the
This perception has gained strength since the implementation of the FCPC award, which is seen to have removed any disadvantages in salaries and wages that may have earlier been faced by government servants. 4 This observation suggests that a close look be given to the past trends in pension payments, but that is not the purpose of this paper, which focuses on the future.
6
Expert Group) projects an annual growth rate of 4.8 per cent during 2000-2010 for all these departments.5
If the estimates of these reports are to be credited, then the future prospect
of Central government finances under the ‘Pension’ head does not seem to be a matter of concern at all, since the projected growth rate of nominal pension payments is lower than the assumed inflation rate. Thus, even if the economy were not to grow at all in real terms, the pension bill to GDP ratio would decline secularly. In such a situation, shifting to a contributory pension system simply does not make sense, since it would involve additional fiscal outgo at a time when the fisc is under considerable strain, with benefits accruing only in the distant future, when fiscal pressures may have already eased.
However, going by the present pension structure, in which most
components of retirement benefits are indexed to price change, it does not seem prudent to place our faith on such projections. The projected outcomes would be possible only under a scenario of declining number of pensioners in future years. However, the trends in employment in the government sectors during last 50 years, and improved life expectancy at age 60, do not support a declining figure for the number of pensioners for many years to come. All of this argues for developing a more appropriate model than used in these studies.
In an earlier study,6 the authors have already developed a rigorous
theoretical model for projecting both the number of pensioners and the pension liabilities from the existing data set available in India.7 Using this model, it was estimated that with an assumed inflation of 6% and an estimated pensioners’ growth of about 2.2%, a realistic assessment of future pension liability for the civil departments of the Central government alone would place the annual nominal growth rate of pensions at 7.6 per cent. The rather substantial discrepancy between this estimate and those made in the other studies suggests that it may not be wise to be overly sanguine about the future course of pension liabilities. This is, however, only a partial picture, since the Civil departments constitute just about 20% of the total pension bill of all departments put together. It would, therefore, be useful to also look into the future pension liability of the other four departments of the Central government, who maintain their separate accounts for the purposes of all transactions including that of payment of salary and pension.
The principal objective of this paper, therefore, is to make a realistic
assessment of the future pension liability of the departments of Railways, Telecommunications, Posts and Defence, and to provide a comprehensive
5 Both these projections are made in nominal terms under an assumed scenario of an annual inflation rate of 6 per cent. 6 Sen and Swain, 2002. 7 One of the main issues raised in Sen and Swain (2002) is the pathetic state of data on retirements, pensioners and pensions in the Indian government, and therefore the need to develop a theoretical model which can address the issues within the constraints posed by the data.
7
projection of total Central government payments under the ‘Pension Accounts’. While doing so, the paper also sheds some light on the dynamics of different kinds of pensioners, and thereby on some considerations that need to be taken into account while framing an alternative pension scheme for government employees.
The growth path of the government pension bill, by its very definition, is
clearly driven by the pattern of growth in two variables: (a) the number of pensioners; and (b) the average pension per pensioner. It is important, therefore, at the first instance to get the correct number of beneficiaries who are likely to be covered under the government pension scheme. The second section in this paper discusses the state of the data relating to the number of pensioners and the need to devise alternative methods for making projections. It presents a model-based projection of the number of pensioners, with validation from past experience.
As per the present system of government pension, a government
employee is entitled to a number of post-retirement benefits in addition to the monthly basic pension. These benefits are always included in the data pertaining to pensions, and thus need to be worked out while computing pensionary liabilities. While the monthly basic pension is applied to all the pensioners cumulated over time, other post retirement benefits accrue to fresh retirees only. Further, the basic monthly pension and other post retirement benefit are different for different category of pensioners. Hence it would not be appropriate to base our estimates on a single average pension figure. The third section of this paper discusses the present system of pension structure and other retirement benefits as it is applied to different categories of pensioners. The specificity of different accounting departments in conformity with present system of Central government pension scheme has also been highlighted in this section.
The projected annual pension liability for each department of Central
government for the period ending 2009-10 is presented in the fourth section. The last section briefly puts the macro economic and fiscal implication of such projection. It would be useful at this stage to mention that our analysis in the following sections are confined to four accounting departments of the central government namely Defence, Railways, Posts and Te lecommunication. The analysis relating to Civil department have been detailed in the earlier paper (ibid, 2002). However, the summary analysis includes the findings on all the five departments, including that on Civil departments, with a view to presenting a comprehensive picture. Projecting the Number of Pensioners
The available information relating to annual retirements and annual
recruitments in the past in the Central government departments and the cumulative number of pensioners along with the category-wise break-up are
8
highly unsatisfactory, to say the least. There are huge data gaps concerning these numbers in virtually all the departments. In the case of the department of Railways, which is the only department to document the number of pensioners and retirees annually for the past several years, there are inconsistencies between the time series on the number of retirees and that on the cumulative number of pensioners, which make their use problematic. The only statistical information available in the published documents with reasonable degree of reliability are the number and distribution of Central government employees across the five accounting departments for past few decades.8 Thus, the past employment growth necessarily has to form the basis to make future projection of the number of pensioners by applying a theoretically justifiable model.
As per the theoretical model developed by the authors,9 the annual rate of
retirement, and the annual number of retirees derived there from, can be computed from the past growth pattern in employment by using certain parametric assumption relating to the average years of service rendered by the government employees and the in-service death rate. Similarly, net accretion made to the stock of pensioners can be estimated from the series of annual retirees by applying parametric values to the attrition rate of pensioners, which in turn is driven by the pattern of retirement. Since the growth pattern in employment observed in the past and nature of retirement vary from department to department, the parameters underlying the projection of the number of annual retirees and pensioners would differ across departments. This paper, therefore, takes into account this inter-departmental diversity and discusses the future growth in the number of pensioners separately for each accounting department.
Before presenting the model, it would first be necessary to define the
different categories of pensioners which are covered under the prevailing pension regime. For the purpose of our projections, pensioners are divided into three categories. The first, and the most important, category comprises of the retired employees, commonly known as service pensioners (SP). The service pensioners (SP) retire from the service either due to superannuation or through voluntary retirement. On an average, a Central government servant, except for those in the Armed forces, renders service of 33 years before superannuating. The time series on SP can be directly computed from the data on annual number of retirees. The second category of pensioners comprises of switch-over family pensioners (SOFP). The SOFP are dependants entitled to receive family pension after the death of the spouse, who were already in receipt of pension. Thus the number of SOFP is a function of SP and their attrition rate. The demographic structure of the government employees along with the varied demographic characteristics for men and women put an increasing number of beneficiaries in the SOFP group. The third category of pensioners is termed as fresh family pensioners (FFP). The FFP are entitled to draw family pension due to in-service death of their spouse. The
8 The employee’s number for the armed forces, category-wise, have been obtained from the Ministry of Defence. 9 Op cit. Sen and Swain (2002)
9
number of FFP can be directly derived from the employees’ strength and the death rate of the age cohort to which the Government employees belong.
The available estimates on future pension liability of Central government
distinguish between service pensioners and family pensioners. However, further segregation of family pensioners into SOFP and the FFP has not been made for the purpose of projection. In our opinion, it is important to distinguish between the two categories of family pensioners due to three compelling reasons. First, the structural dynamics of each of these categories are different as should be evident from their very definition. Second, these two categories of family pensioners are driven by different demographic factors with different average life expectancy and it is important to treat them separately for the purpose of projections. Third, the pension entitlements of the two categories of family pensioners are different and clubbing them together can lead to serious projection errors.
While the available statistics on the number of pensioners is in itself of
questionable quality, there is absolutely no clue about the number of SOFP and FFP. Some rough estimates of family pensioners and service pensioners covered by each of the accounting departments are available in the published documents, such as Pay Commission Reports, annual publications of Ministries and Departments and report of the Working Groups. In the absence of any other reliable estimates, the exercise undertaken in this paper takes the base year (1999-2000) statistics from these reports. The first step in projecting the number of pensioners has been to break up the total number of pensioners of the base year (1999-00) to three categories of pensioners namely SP, SOFP and FFP. The available data on family pensioners for the year 1999-2000 as obtained from various reports and published documents have been taken as the basis of total family pensioners. The estimation of FFPs has been made for the year 1999-2000 by building up a time series of FFPs prior to 1999-2000 from the past data on employment of respective departments. The number of FFP has been netted out from the total family pensioners to obtain the number of SOFP for the base year10.
After obtaining the base year figure for the three categories of pensioners,
the projection of total number of pensioners for the period 2001-2010 has been made in two steps. First, the annual number of retirees has been estimated for each department on the basis of the following equations:11
Rt = rt*( Nt) (1) rt = (1 - λλ .δδ )(g + δδ )/{(1 + g)λλ - (1 - λλ .δδ )} (2) at = rt + δδ (3)
10 Pensioners’ number under Railways Accounts does not provide any category-wise break up. The base year figure on total number of pensioners has been broken down into the three categories of pensioners by using past employment statistics. 11 See Sen & Swain (2002) for derivation of these equations from the structural model.
10
where:
Rt = Number of retirees in year t Nt = Number of employees in year t rt = percentage of employees retired in year t λ = The average length (years) of service in government g = average growth rate of the number of government employees over the period (t -
λ) to t δ = Average annual death rate for government employees, given the age distribution. at = attrition rate of government employees in year t.
The model-based estimates of annual number of retirees, as derived from
the equation, refutes the conventional wisdom based on rule of thumb principle that the number of employees retiring annually is a constant ratio of the number of employees and that this ratio is the inverse of the average years of service the employee puts in. It is only in the limiting case of g = 0% with value of λ = 33 years and δ = 0.0032 that rt = 2.68%. Addition of annual average in-service death rate of 0.32% to the annual retirement rate in the limiting case makes the annual attrition rate of government employees equal to 3%, i.e. the inverse of 33 years.
It would be useful to look at some indicative figures of retirement and attrition
rates given by the model at different rates of growth of the number of government employees as presented in Table-1. These figures have been derived under the assumptions that the average length of service (λ) is around 33 years and the average annual rate of in-service deaths (δ) is 0.32%.12 As may be seen from the table, the 3% attrition rate obtains only when the rate of growth of the number of government employees over the past 33 years has been zero. If the growth rate had been positive, the retirement and attrition rates will be progressively lower. On the other hand, if there has been a process of downsizing in the past – i.e. the growth rate of employment is negative – the retirement and attrition rates tend to rise sharply. Thus, it becomes virtually impossible to specify a unique retirement or attrition rate independently of the past behaviour of government employment.
In interpreting this table, it must be borne in mind that the retirement and
total attrition rates relate to a particular year, while the growth rate of government employment is the average over the previous 33 years. Thus the model is based on one critical assumption – that government employment has followed a steady-state growth path for a long period in the past. This is clearly not a realistic assumption and deviation from this assumption would have different implications on the behaviour of future retirement rate. On the whole, however, the divergences from a steady-state would show up as annual variations in the retirement rate around a mean, which would be determined by the average
12 This value of δ has been obtained from the age-specific death rates for urban males in India for
the age-group of 23 to 58 years.
11
growth rate over the concerned period. Thus the annual variations would tend to smoothen out over time.
Table 1: Retirement and Attrition Rates at Different Rates of Growth
Rate of growth of government employment Retirement Rate r(2)Attrition Rate
-3.0% 4.51% 4.83% -2.5% 4.20% 4.52% -2.0% 3.92% 4.24% -1.5% 3.65% 3.97% -1.0% 3.41% 3.73% -0.5% 3.41% 3.73% 0.0% 2.68% 3.00% 0.5% 2.55% 2.87% 1.0% 2.36% 2.68% 1.5% 2.17% 2.49% 2.0% 1.99% 2.31% 2.5% 1.82% 2.14% 3.0% 1.66% 1.98% 3.5% 1.51% 1.83% 4.0% 1.38% 1.70%
Note: Columns 2 and 3 are computed from equations (2) and (3)
respectively with assumed parameter values of δ = 0.0032 and λ = 33.33.
It must further be noted that the parametric value of λ and δ used in
deriving the numbers shown in the table may not be universally applicable. These values mostly hold good for civilian government employees in the Central Government in India. However, the service condition, pattern and structure of employment, age cohort of the fresh recruits, which have bearing on λ and δ, differ from department to department and these initial conditions would have implication for the value assigned to λ and δ. Accordingly, the retirement rate of each accounting department would have to be computed independently by assigning different values to λ and δ on a case-to-case basis in order to avoid projection bias. The following paragraphs explain the estimation of department-wise retirement rate and annual retirees over a period of time. Retirements in Railways
The department of Railways accounts for more than half of the total civilian
employees in the Central government. The employment statistics of Railways are available for an extended period of time starting from the pre-independence period. Observation of the time series on employment indicates fairly stable growth in post-1965 employment, which is the relevant period for this exercise. A slight down turn is observed in the employment after the year 1995, which may be
12
the effect of the downsizing effort of the government. Our estimation of the model in case of Railways, therefore, is based on the assumption of steady-state growth path of employment in the past.
Assigning values to the parameters λ and δ in case of Railways would
require an understanding of the working environment of the employees, their job location, etc. Most of the railway employees belonging to the group C and group D categories13 render their services either on-board inside the train or in remote location, hardly having any provision for catering to the basic needs like health and education facilities for their family members. In most cases, they are compelled to stay away from their families. Hence, there is reason to believe that it is in the interest of the employees to opt for voluntary retirement much before attainment of the age of superannuating. In this exercises, we assume that about 25% of the employees would opt for voluntary retirement after completion of 20 years of service, the minimum years required to qualify for pension. The balance 75% would have a normal service career of 33 years. Thus, on average a railway employee would render service for a period of about 30 years.
Early retirement age of the railway employee would imply relatively
younger age cohort for these employees. Accordingly the average death rate would be less than normal rate of 0.3%. For our estimation a death rate of 0.28% has been assumed. After assigning value to the death rate and average year of service, the rate of retirement for the employees in Railways department have been computed as presented in the Table–2. As can be seen the result is consistent with the model findings depicted in Table 1. Higher employment growth rate in the preceding 30 years would lead to lower retirement rate and vice-versa.
Table 2: Employment and Retirement in Railways
Number of Quinquennial Growth rate Rate of Annual
Year Employees Growth rates Over 30 years Retirement Retirees (est) 1960 1142776 1961 1148287 1962 1162505 1963 1196102 1964 1251296 1965 1296018 2.5% 3.1% 2.0% 25729 1966 1326948 3.2% 2.0% 26133 1967 1340849 3.2% 2.0% 26211 1968 1345650 3.3% 1.9% 25782 1969 1338685 3.3% 1.9% 25893 1970 1344699 0.7% 3.2% 1.9% 26105 1971 1359571 3.2% 2.0% 26735 1972 1378931 3.1% 2.0% 27560 1973 1411000 2.8% 2.1% 29277
13 Groups C &D together constitute more than 98% of the railways employees as per the Railways records.
13
1974 1432000 2.6% 2.1% 30701 1975 1441000 1.4% 2.4% 2.2% 31994 1976 1457000 2.3% 2.3% 32834 1977 1463983 2.2% 2.3% 33483 1978 1471000 2.1% 2.3% 34338 1979 1497000 1.8% 2.5% 36720 1980 1553000 1.5% 2.0% 2.4% 36777 1981 1572000 1.8% 2.4% 38119 1982 1575000 1.8% 2.4% 38322 1983 1583000 1.8% 2.4% 38480 1984 1592000 1.7% 2.5% 39351 1985 1603000 0.6% 1.7% 2.5% 39992 1986 1613000 1.5% 2.5% 40947 1987 1611000 1.4% 2.6% 41526 1988 1618000 1.3% 2.6% 42683 1989 1626000 1.2% 2.7% 43471 1990 1647000 0.5% 1.2% 2.7% 43800 1991 1652000 1.2% 2.7% 43971 1992 1654000 1.2% 2.7% 44268 1993 1645000 1.1% 2.7% 44769 1994 1625000 0.9% 2.8% 45469 1995 1602000 -0.6% 0.7% 2.9% 45878 1996 1586000 0.6% 2.9% 46121 1997 1584000 0.6% 2.9% 46303 1998 1579000 0.5% 2.9% 46295 1999 1578000 0.5% 2.9% 46173 2000 1577000 -0.3% 0.5% 2.9% 46247
The estimated annual number of retirees increases at a steady rate
gradually. This is due to the assumption of the steady state growth in employment in the past. In reality there has been some fluctuation in the growth rate of employees in the past. The impact of this has to be reflected in the actual number of annual retirees. The Graph 3 indicates a comparative picture of actual retirement rate and predicted retirement rate in the Railways since the year 1975. As can be observed, the discrepancy between the estimated rate and the actual rate of retirement over a period of 25 years is not very significant. The fluctuation of actual series around the theoretical series is negligible on average. Thus, the model finding, i.e. the predicted rate of retirements in case of railways employees, appears to be reliable and application of this model to make projection of annual retirements in future is expected to provide reasonably accurate estimates.
14
Graph 3: Rate of Retirement in Railways
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
1974
-75
1975
-76
1976
-77
1977
-78
1978
-79
1979
-80
1980
-81
1981
-82
1982
-83
1983
-84
1984
-85
1985
-86
1986
-87
1987
-88
1988
-89
1989
-90
1990
-91
1991
-92
1992
-93
1993
-94
1994
-95
1995
-96
1996
-97
1997
-98
1998
-99
1999
-2000
Years
Re
tire
me
nt
Ra
te
Actual Estimated
For estimating the retirement rate of railways employees during the period
2001-2010, it is assumed that the employment figure for future years would remain same at the base year (1999-2000) level. Table 3 indicates the projected number of retirees in Railways during 2001-10. As can be seen, the rate of retirement increases gradually corresponding to a falling rate employment growth during this period.
Table 3: Projected Number of Retirees in Railways
Year Number of Rate of No. Employees Retirement Retirees
2001 1577000 2.9% 45970 2002 1577000 2.9% 46241 2003 1577000 3.0% 46668 2004 1577000 3.0% 46930 2005 1577000 3.0% 47038 2006 1577000 3.0% 47222 2007 1577000 3.0% 47298 2008 1577000 3.0% 47373 2009 1577000 3.0% 47627 2010 1577000 3.0% 48002
15
Retirements in Defence
Next to Railways, Defence is the second largest employer in the Central government. The employees’ strength in all the three wings of armed forces taken together touches more than 1.2 million. The service condition of the personnel in the armed forces, the age cohort of the employees and the pattern of retirement vary significantly from that of civil employees. Further, within the armed forces itself, service conditions and retirement pattern varies between Commissioned Officers (CO) and Personnel Below Officers Rank (PBOR). Application of the model (equation 2 and 3) in case of Defence would therefore require careful consideration of the values to be assigned to λ and δ.
In this case, there are two distinct series of employment statistics available
separately for PBOR and CO. These two categories of personnel represent two distinct age cohorts and are governed by different service conditions. The officers, in general, render services for a much longer period than the PBOR. But average years of service rendered by CO would be less than that of civilian employees due to younger age of retirement for most of the service officers compared to their civilian counterparts. In our exercises, it is assumed that the period of service rendered by an officer on average would be 30 years. The average length of service for the PBOR has been estimated to be in the range of 20 years.
Early age of induction in the defence services and early age of retirement,
particularly in case of PBOR, who constitute about 85% of the total employment in armed forces, imply a younger age cohort, on average, for defence employees compared to that of civilians. In the absence of death by violence during war or other such events, the younger age cohort should imply a significantly lower ‘normal’ in-service death rate. However, the empirical evidence indicates an average in-service death rate of about 1.2% among the defence personnel.14 This could be attributed to the death caused by fighting against internal insurgency and combating terrorism, among others. Adoption of the model to defence employees, therefore, assumes in-service death rate of 1.2%, which is consistent with the observed statistics.15 Estimation of retirement rate in the model takes into account only those, who retire from the service after attaining the age of superannuation. However, in defence services there are a large number of retirements categorized under voluntary retirement, invalid retirement, compulsory retirement. In the past 10
14 The death in–service among defence personnel for the last 10 years has been obtained from the Defence Pension Accounts Office. 15 The value assigned to the death rate is same for PBOR and CO since separate figures for the two categories are not available.
16
years these groups constitute about 1.5% of the total employees. In our exercises we group them together under ‘Voluntary Retirement’ and estimate a series on voluntary retirements separately as a fixed ratio of employment for each year.
As explained above, the estimation of retirement rate for defence employees has been made in three steps. Retirement rate and number of retirees have been estimated separately for PBOR and CO by assigning different values of λ and δ in equation 2 of the model as has already been explained. The retirements occurring due to reason other than superannuation, termed as voluntary retirement in this case, has been estimated separately by assigning a fixed ratio (1.5%) to the employment data. The three categories of retirees are combined together to arrive at the total number of annual retirements in Defence department.
Table 4 and Table 5 below contain the predicted rate of retirement and
predicted number of retirees’ respectively for PBOR and CO. The retirement rate for PBOR mostly remains at around 3.6% during the period 1990 to 2000. The retirement of CO during the same period is estimated to be within the range of 2.0 to 2.8%. Combining CO and PBOR the total number of defence personnel retiring annually due to superannuartion has been computed to be around 3.5%. In absolute numbers, more than 44000 army personnel are superannuating every year.
17
Table 4: Employment and Retirement of PBOR in Defence
Number of Rate of growth Rate of Annual
Year PBOR over 20 years Retirement Retirees
1960 346832 3.3% 11445
1961 398857 2.1% 3.3% 13289
1962 462701 2.8% 3.1% 14399
1963 536974 3.5% 2.9% 15537
1964 623391 4.3% 2.7% 16713
1965 723953 5.0% 2.5% 17933
1966 840990 5.8% 2.3% 19199
1967 850173 5.8% 2.3% 19344
1968 859748 5.6% 2.3% 19985
1969 869743 5.7% 2.3% 20142
1970 880191 5.7% 2.3% 20304 1971 891127 5.8% 2.3% 20470 1972 894562 5.7% 2.3% 20562 1973 898030 5.7% 2.3% 20655 1974 901531 5.7% 2.3% 20749 1975 905066 5.7% 2.3% 20843 1976 908635 5.7% 2.3% 20938 1977 914770 5.7% 2.3% 21058 1978 920948 5.7% 2.3% 21179 1979 927168 5.0% 2.5% 22979 1980 933430 4.3% 2.7% 24883 1981 939735 4.4% 2.7% 24961 1982 964076 3.7% 2.8% 27339 1983 989074 3.1% 3.0% 29883 1984 1014749 2.5% 3.2% 32580 1985 1041118 1.8% 3.4% 35393 1986 1068201 1.2% 3.6% 38238 1987 1084938 1.2% 3.6% 38767 1988 1102024 1.2% 3.6% 39310 1989 1119472 1.3% 3.6% 39867 1990 1137294 1.3% 3.6% 40441 1991 1155503 1.3% 3.6% 41031 1992 1144556 1.2% 3.6% 40856 1993 1132727 1.2% 3.6% 40656 1994 1118121 1.1% 3.6% 40385 1995 1109443 1.0% 3.6% 40243 1996 1130903 1.1% 3.6% 40794 1997 1145610 1.1% 3.6% 41229 1998 1159232 1.2% 3.6% 41640 1999 1167617 1.2% 3.6% 41933 2000 1172623 1.1% 3.6% 42152
18
Table 5: Employment and Retirement of CO in Defence
No. Of Officers Rate of growth Rate of retirmt Annual
Year Over 30 years (model) Retirees
1960 15879 1.5% 238
1961 18261 1.5% 274
1962 21084 1.5% 316
1963 24355 1.5% 365
1964 28149 1.5% 422
1965 32549 1.5% 488
1966 37653 1.5% 565
1967 38198 1.5% 573
1968 38760 1.5% 581
1969 39340 1.5% 590
1970 39939 1.5% 599 1971 40558 4.1% 1.5% 623 1972 40812 4.1% 1.5% 626 1973 41080 4.1% 1.5% 630 1974 41360 4.1% 1.5% 633 1975 41654 4.1% 1.5% 637 1976 41964 4.2% 1.5% 640 1977 42859 4.2% 1.5% 647 1978 43774 4.3% 1.5% 653 1979 44709 4.2% 1.5% 680 1980 45664 4.2% 1.5% 687 1981 46641 4.3% 1.5% 693 1982 47412 4.3% 1.5% 699 1983 48199 4.4% 1.5% 705 1984 49004 4.4% 1.5% 711 1985 49827 4.4% 1.4% 717 1986 50668 4.5% 1.4% 723 1987 52170 4.6% 1.4% 732 1988 53722 4.7% 1.4% 741 1989 55328 4.7% 1.4% 750 1990 56987 4.4% 1.5% 835 1991 58703 4.0% 1.6% 929 1992 58969 3.5% 1.7% 1030 1993 57415 2.9% 2.0% 1134 1994 57535 2.4% 2.2% 1262 1995 56631 1.9% 2.5% 1405 1996 56781 1.4% 2.8% 1582 1997 56439 1.3% 2.8% 1600 1998 56785 1.3% 2.9% 1621 1999 58384 1.3% 2.8% 1649 2000 59299 1.3% 2.8% 1674
19
In order to validate our model as applied to Defence Accounts, it would be desirable to plot the predicted retirement rate against actual rate of retirement, which has been computed from the number of PPO’s issued under Defence Accounts available for the last 10 years. Graph 4 captures the estimated retirement vis-à-vis the actual retirement rate for the last decade. As can be seen, the actual rate of retirement has fluctuated around the predicted line during the 1990s in a symmetrical fashion, except for the spike in the years 1996-97 and 1997-98. Past employment data does not support such a large figure of actual retirement in defence in the year 1997 and 1998. It is worth mentioning here that in the absence of data on actual number of annual retirees, the number of PPOs issued annually has been used as proxy for the annual retirement. There is a possibility that the number of PPOs may overshoot the number of actual retirements due to issuance of some revised pension orders in that year. Thus, it seems reasonable to use the model for projecting the future retirement in Defence department.
Graph 4 : Retirement Rate in Defence
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
5.0%
5.5%
1990-91 1991-92 1992-93 1993-94 1994-95 1995-96 1996-97 1997-98 1998-99 1999-00
Estimated Actual
As in the case of other departments, the future employment in Defence is assumed to remain at the base year level during the 10 years period starting 2001. The model-based projection of the annual retirements is estimated to increase gradually to more than 46000 from the present level of about 43000. In addition to this, about 1.5% of the employees retire from the service voluntarily. Table 6 summarises the aggregate retirement position projected for the defence employees for the period 2001 to 2010.
20
Table 6: Projected Employment and Retirement in Defence
Year No. of Employees Retirees VR Total (PBOR+CO) Superannuating Retirees
2001 1231922 43955 18479 62434 2002 1231922 44358 18479 62837 2003 1231922 44744 18479 63223 2004 1231922 45107 18479 63586 2005 1231922 45443 18479 63922 2006 1231922 45744 18479 64222 2007 1231922 45930 18479 64409 2008 1231922 46100 18479 64579 2009 1231922 46249 18479 64728 2010 1231922 46373 18479 64852
Retirements in Post Application of the model to estimate retirement rate in the Department of Post assumes similar service condition for postal employees as that of civil departments. Therefore, the values assigned to λ and δ are 33 years and 0.32 per cent respectively. The past employment behavior in regard to Post follows a fairly stable growth path up to the year 1984, and thereafter there is a slow and steady decline in the number of employees. It would be worth mentioning here that the employment figure presented in the Table 7 is confined only to the departmental staff. However, the Department of Post employs an equivalent number of personnel as extra-departmental employees, and these extra-departmental staff are not entitled to receive pension under the prevailing pension scheme. Our exercise assumes the present position concerning extra departmental employees to remain unchanged, and hence applies the model to the employees net of extra departmental ones for the purpose of theoretical projection. The predicted rates of retirement during the past thirty years have been presented in Table 7. As can be seen, there is slow and gradual increase in the retirement rate during last 3 decades. The predicted pattern of retirement is consistent with the past employment behaviour, which exhibited a declining long run growth pattern. While it is important to establish the robustness of our estimation by placing the predicted value across the observed value of the retirement rate, non-availability of actual retirement data in case of the Department of Post prevents us from establishing such validation.
21
Table 7: Employment and Retirement in Post Number of Growth rate Rate of Number of
Year Employees Over 33 years Retirement Retirees 1960 157117 1961 160000 1962 165400 1963 170982 1964 176753 1965 182718 1966 188885 1967 195260 1968 201850 1969 208662 1970 215705 1971 222985 2.3% 2.1% 4597 1972 227955 2.3% 2.1% 4699 1973 233036 2.3% 2.1% 4804 1974 238231 2.3% 2.1% 4911 1975 243541 2.3% 2.1% 5021 1976 248970 2.3% 2.1% 5133 1977 254519 2.3% 2.1% 5247 1978 260192 2.3% 2.1% 5364 1979 265992 2.3% 2.1% 5484 1980 271921 2.4% 2.0% 5545 1981 278000 2.4% 2.0% 5650 1982 293000 2.5% 2.0% 5823 1983 309000 2.6% 1.9% 6003 1984 310000 2.6% 2.0% 6074 1985 308053 2.5% 2.0% 6124 1986 298707 2.3% 2.0% 6119 1987 291957 2.2% 2.1% 6133 1988 291478 2.2% 2.1% 6195 1989 288421 2.1% 2.2% 6238 1990 286822 2.0% 2.2% 6294 1991 293225 2.0% 2.2% 6413 1992 291527 1.9% 2.2% 6469 1993 290035 1.9% 2.3% 6529 1994 290197 1.8% 2.3% 6608 1995 288552 1.7% 2.3% 6741 1996 288498 1.6% 2.4% 6892 1997 286378 1.5% 2.5% 7033 1998 293979 1.5% 2.5% 7255 1999 293072 1.3% 2.5% 7419 2000 294301 1.3% 2.6% 7607
22
However, the Working Group report contains some figures relating to the projected number of pensioners in the department of post for the years 2001 to 2010. The report also provides the basis for the estimation of such projected numbers. In order to satisfy ourselves with the model result, it was necessary to calculate backward the annual number of retirees from the projected number of pensioners. The implicit retirement rate derived therefrom has been compared to the model-based projection of the rate of retirement for the current decade as shown in the Graph 5 below. As can be noticed our model-based projection almost converges with the Working Group numbers except for the years 2000-01 and 2002-03. It is therefore reasonably accurate to apply the model for purpose of our projection. Table 8 below presents the retirement numbers in absolute terms.
Graph 5
Projected Retirement Rate in Post
1.5%
2.0%
2.5%
3.0%
3.5%
2000-01
2001-02
2002-03
2003-04
2004-05
2005-06
2006-07
2007-08
2008-09
2009-10
Year
Ret
irem
ent
Rat
e
Model
Working group
Table 8: Projected Retirement in Post Year Number of Rate of No.
Employees Retirement Retirees 2001 294301 2.65% 7796 2002 294301 2.72% 7996 2003 294301 2.79% 8210 2004 294301 2.87% 8442 2005 294301 2.92% 8608 2006 294301 2.98% 8785 2007 294301 3.05% 8975 2008 294301 3.12% 9181 2009 294301 3.20% 9408 2010 294301 3.28% 9659
23
Retirements in Telecommunication
The terms and conditions of the service, average age cohort, demographic structure and pattern of retirement for employees belonging to the Department of Telecommunication are almost similar to that of Central government civilian employees. Hence it is reasonable to assign same value to λ and δ as that in case of civil departments. The average years of service rendered by an employee in the telecommunication department is, therefore, assumed to be 33 years and the in-service death rate assumes a value of 0.32%. However, the employment pattern in the Department of Telecommunication in the past indicates two distinct time phases. The observed employment statistics exhibited a very high growth, measuring 5.6% annually, during 1971 to 1981 as compared to a growth rate of around 2.5% both before 1971 and after 1981. The base data are presented in the Table 9. This growth pattern has obvious implications for retirements during the decade starting 2001. Therefore it may not be desirable to assume a steady- state growth path of employment in the case of Telecommunication.
Table 9: Employment and Retirement in Telecommunication
Year Number of Decadal Growth rate Rate of Annual
Employees Growth Over 33 years Retirement Retirees 1960 124182 1961 127310 2.5% 1962 133166 1963 139292 1964 145699 1965 152402 1966 159255 1967 160768 1968 162295 1969 163837 1970 165393 1971 167000 2.8% 1972 176414 1973 186358 1974 196863 1.2% 2323 1975 207960 1.2% 2454 1976 219683 1.2% 2592 1977 232067 1.2% 2738 1978 245148 1.2% 2893 1979 258967 1.2% 3056 1980 273565 1.2% 3228 1981 287523 5.6% 1.2% 3393 1982 295396 1.2% 3486 1983 302498 1.2% 3569 1984 310041 3.5% 1.2% 3658 1985 317875 1.2% 3751 1986 327067 1.2% 3890 1987 335704 1.2% 3995
24
1988 344332 1.2% 4138 1989 351987 1.2% 4251 1990 358943 1.2% 4343 1991 376703 2.7% 3.6% 1.2% 4597 1992 374608 1.2% 4593 1993 386987 1.2% 4760 1994 397913 1.2% 4930 1995 422722 1.3% 5559 1996 424456 1.3% 5626 1997 432682 1.3% 5777 1998 429015 1.3% 5770 1999 428069 1.4% 5801 2000 424983 3.0% 1.4% 5802
The methodology for projecting future retirements in this case implicitly
adopts the model described above, but computes the retirement numbers in a step-wise manner on a year-to-year basis. The predicted retirement rate and the corresponding number of annual retirees up to the year 2000 are presented in Table 9 above. Unfortunately, the actual number of annual retirees in the past for the telecommunication department is not available. Hence validation of the methodology by tracing a comparison between the actual values and predicted values has not been possible in this case. More relevant for the purpose of our study, however, is the projected rate of retirement and the corresponding number of retirees for the future years, which are presented in Table 10.
Table 10: Projected Retirements in Telecommunication
Number of Rate of No.
Year Employees Retirement Retirees 2001 426606 1.4% 5868 2002 426606 1.4% 6015 2003 426606 1.5% 6186 2004 426606 1.5% 6356 2005 426606 2.5% 10645 2006 426606 2.6% 11220 2007 426606 2.8% 11860 2008 426606 2.9% 12529 2009 426606 3.1% 13225 2010 426606 3.3% 13950 2011 426606 3.5% 14745 2012 426606 3.6% 15568 2013 426606 3.9% 16437 2014 426606 4.1% 17354 2015 426606 2.1% 9032 2016 426606 2.2% 9211 2017 426606 2.2% 9395 2018 426606 2.2% 9590 2019 426606 2.3% 9781 2020 426606 2.3% 9993
25
As has been explained earlier, these projections have been made by applying the model without assuming a steady-state time profile of past employment growth in telecommunication. This was necessary because, unlike in the case of other departments, projection of the annual retirees under the assumption of a steady-state growth path for telecommunication would lead to obvious downward bias in the medium term, although the longer-term projections would probably continue to remain more or less valid. The implication of the two alternative computational methodologies on the future rates of retirement is highlighted in the Graph 5. It can be clearly noted that the assumption of steady state growth in the past employment numbers for the te lecommunication underestimates the rate of retirement for the period during 2004 to 2014. Over the longer term, however, the rates converge. Since the objective of this paper is to project the pension bill of the central government for the years 2001 to 2010, it is clearly appropriate to apply the model variant in this case.
As can be seen from the graph, the retirement rate is projected to increase
from the present level of 1.4% to 2.5% in the year 2004-05. Thereafter, the retirement rate keeps increasing sharply till it reaches the level of 4.1% in the year 2014 and then suddenly drops to the level of 2.1% in the year 2015 and gradually increases thereafter.
Graph 5: Rate of Retirement in Telecommunication
0.00%
0.50%
1.00%
1.50%
2.00%
2.50%
3.00%
3.50%
4.00%
4.50%
2000
-01
2001
-02
2002
-03
2003
-04
2004
-05
2005
-06
2006
-07
2007
-08
2008
-09
2009
-10
2010
-11
2011
-12
2012
-13
2013
-14
2014
-15
2015
-16
2016
-17
2017
-18
2018
-19
2019
-20
Year
Rat
e of
Ret
irem
ent
Different time profile of growth
Steady state growth
26
Projecting the Number of Pensioners After estimating the number of annual retirees as described above, the net
accretion to the base year’s stock of pensioners can be determined by the number of retirees (R) during the year, which is the gross accretion to the stock, less the number of deaths among the existing service pensioners, which is the attrition from the stock. However, the task is not so easy, since application of a constant ratio average death rate to the stock of pensioners may lead to gross statistical error arising due to adoption of the same simple rule of thumb principle in this case also. The assumed attrition rate of the pensioners referred to in various extant reports has been computed as the inverse of the number of years a pensioner is expected to survive. But this principle would be valid only in the limiting case where: (a) number of any given category of pensioners has remained constant over a period of time; and (b) the base data on the number of pensioners are in conformity with the past employment figure16. Past trends in government employment and the pension bill are clearly not supportive of these two presumptions. The present exercise therefore applies the following mathematical relationship17 to compute the net accretion to the number of pensioners in different categories.
∆∆SPt = Rt – Rt-γγ 1 (4) ∆∆SOFPt = (1 – δδ 1) (Rt-γγ 1 – Rt-γγ 1-γγ 2) (5)
∆∆FFPt = δδ (1 – δδ 2) (Nt-1 – Nt-1-γγ 3) (6) Where: ∆∆SPt, ∆∆SOFPt and ∆∆ FFPt are net accretion to the stock of SP, SOFP and FFP respectively. γγ1 = average years of survival after retirement. It is assumed for simplicity that the entire cohort of retirees in a given year dies simultaneously after γ1 years, δδ 1 = percentage of service pensioners with no dependents γγ2 = average years of survival of SOFP. It is assumed that the cohort of dependents who receive family pensions on the death of the pensioner in a given year die simultaneously γ2 years later δδ 2 = percentage of government employees dying in service without dependants γγ3 = average years of survival of FFP. As in the other cases, death among fresh family pensioners is also assumed to be on a cohort basis.
16 There are reasons to believe that base year figure of number of pensioners does not take into account all the past retirees under pension cover. 17 For derivation see Sen and Swain, 2002.
27
Equations (4), (5) and (6), therefore, represent the annual net accretions to the number of SPs, SOFPs and FFPs respectively. It can be seen that most of the data, particularly for R and N over an extended period of time, required for making future projections have already been presented in earlier tables spanning Table 2 to Table 10. The basis of arriving at such employment and retirement numbers has also been explained. However, before entering into projections on number of pensioners, it is necessary to provide values for the parameters represented by the various δs and γs. Starting with δ, the average in-service death rate among government employees, the estimate of 0.32% (1.2% for Defence) used earlier continues to remain valid. As far as δ1 and δ2 are concerned, it is assumed that all government employees and pensioners have some dependants at their time of death, so that both these parameters can be taken to be zero. This is of course a somewhat extreme assumption, but it is probably not too far off the mark, at least as far as δ2 is concerned.18 Nevertheless, the possibility exists that the rate of accretion may be marginally over-stated. The γ parameters are clearly determined by the assumed longevity of the different categories of pensioners. According to the life tables, the average life expectancy of a male at the retirement age of 58 years is 18 years, i.e. up to 76 years. The female life expectancy is 3 years longer; but with women representing only 11% of government employees, it makes only a marginal difference. Therefore, γ1 has been taken to be 18 years as against 20 years normally presumed in various published documents. In the case of SOFPs, it is assumed that all government employees are male and that the average difference in the age at marriage between men and women is 5 years. This, coupled with a 3 years longer life expectancy of women, yields a value of γ2 of 8 years as compared to the Working Group assumption of 10 years. The largest discrepancy, however, occurs in the case of fresh family pensioners (FFPs). As mentioned earlier, the average age for in-service deaths in government is 54 years, which, taken with the difference in the age at marriage, implies that the average age of a fresh family pensioner would be 49 years. Since the average life expectancy of women in that age group is 79 years, it implies that a FFP lasts for at least 30 years and not 25 years as assumed by the Working Group. The assumed longevity of different categories of pensioners in the defence department would be altogether different from that of civilians, since defence pensioners would comprise a much younger age cohort. The average retirement age of a CO would be less than 58 years and is assumed to be about 52 years. Thus, an average longevity of 26 years has been assumed for army officers. However, the largest chunk of the pensioners belongs to the PBOR, who retire at a much younger age of about 40 years and an average life expectancy of a 18 In the case of service pensioners, however, there is certain ly a fair possibility that there may be no dependants at the time of death, since the probability of pre-decease of the spouse during the relevant age of the government employee, ranging between 23 to 76 years, is quite significant. In addition, at the age of death of 76 years, there may not be any dependent children.
28
PBOR has been assumed to be 36 years. So far as γ2 is concerned, i.e. the average life expectancy of a SOFP in defence, we assign the same value of 8 years as in the case of other departments. Finally, it needs to be reiterated that all our projections depend upon the assumption made about the future time path of the number of government employees. For simplicity it has been assumed that the number of government employees in all categories will be held constant at the base year level for the next ten years. The projections are no doubt sensitive to this assumption, and alternative numbers can be generated on the basis of different assumptions. It should be noted that the model does not in itself generate the stock of pensioners in any given year. It generates only the net addition to the stock of pensioners in each year by calculating the annual accretion to and attrition from a given stock. In order to utilize the model results, therefore, some base year estimate of the stock is necessary on which these annual changes can be applied. As it turns out, the only year for which this data are available is 1999-2000, during which a special survey was undertaken. Since there is no earlier data on the stock, there is no possible way of checking either the consistency between the stock and the flows or on the accuracy and veracity of the base year stock figures. Thus, the projections are sensitive to this one point base figure.
On the basis of the above assumptions and assigned parametric values, category-wise projections of pensioners have been made for the period up to 2009-10 for each department. These are presented in Tables 11 to 14 below.
Table 11: Projected Number of Pensioners in Railways
Years Service SOFP FFP Total
1999-2000 657465 245075 115639 1018178
2000-01 667214 255903 116247 1039364 2001-02 677187 266575 116802 1060564 2002-03 686266 277959 117267 1081491 2003-04 695416 288721 117673 1101810 2004-05 704523 297923 118054 1120500 2005-06 712952 306573 118390 1137915 2006-07 720823 314572 118706 1154101 2007-08 727818 322685 119003 1169506 2008-09 734506 330728 119227 1184461 2009-10 740449 339073 119294 1198817 CAGR 1.2% 3.3% 0.3% 1.6%
29
Table 12: Projected Number of Pensioners in Defence
Years Service SOFP FFP Total
1999-2000* 1433428 249174 273641 1956243 2000-01 1475206 257872 284165 2017244 2001-02 1515785 268108 294671 2078564 2002-03 1556210 278817 305160 2140186 2003-04 1595900 289115 315630 2200645 2004-05 1635217 298246 326082 2259546 2005-06 1674056 307036 335992 2317084 2006-07 1712860 314898 345277 2373035 2007-08 1751685 321493 353796 2426973 2008-09 1790506 326743 361424 2478673 2009-10 1829297 330545 368015 2527857 CAGR 2.5% 2.9% 3.0% 2.6%
Table 13: Projected Number of Pensioners in Post
Years Service SOFP FFP Total
1999-2000 112678 47234 12110 172022 2000-01 114471 48175 12338 174984 2001-02 116393 49052 12534 177979 2002-03 118479 49807 12714 181000 2003-04 120803 50457 12876 184135 2004-05 123278 51107 13021 187405 2005-06 125867 51695 13148 190710 2006-07 128603 52166 13258 194027 2007-08 131491 52576 13348 197415 2008-09 134486 52971 13420 200877 2009-10 137676 52971 13420 204066 CAGR 2.0% 1.2% 1.0% 1.7%
30
Table 14: Projected Number of Pensioners in Telecom
Years Service SOFP FFP Total
1999-2000 62250 20750 17000 100000 2000-01 64548 21866 17831 104245 2001-02 66905 22932 18631 108468 2002-03 69340 23944 19400 112684 2003-04 71807 24941 20135 116883 2004-05 78457 25880 20835 125172 2005-06 85539 26790 21497 133826 2006-07 93148 27648 22120 142916 2007-08 101333 28506 22700 152539 2008-09 109960 29534 23237 162731 2009-10 119318 30468 23727 173512
CAGR 6.7% 6.0% 3.4% 5.7%
Although the category-wise projections are essential for estimating the future pension liabilities of the government, and indeed for any forward-looking exercise on the behaviour of pensions and pensioners, it is also useful to know the total number of pensioners that could exist at any given point in time for monitoring and control purposes.19 Table 15 provides the summary position relating to the projection of the total number of pensioners annually for the years 2001-2010 comprising all departments, including especially the civilian departments, which have otherwise not been examined in detail in this paper.20 The data relating to the future projections of the number of pensioners reveals that the total number of pension accounts are projected to rise from about 3.83 million in 1999-2000 to about 4.83 million by 2009-10, implying a 2.3% annual rate of increase increase in the total, albeit with interdepartmental variations. The highest annual average growth has been projected for the department of telecommunication. This can be attributed primarily to the past employment pattern in this sector, particularly during the period 1971 to 1981, as has been already discussed. This high level of growth would continue to remain for quite some time even beyond 2010. However, pensioners in Te lecommunication constitute less than 4% of the total Central government pensioners, and hence have negligible impact on the total number. The most dominant segment in the entire central government pension system is the Department of Defence, comprising more than half of the total pensioners. The estimated growth path for defence pensioners is relatively high at 2.6% mainly due to the very fact that defence personnel retire at a younger age.
19 The data that should readily be available from the pension disbursing agencies, such as banks and treasuries, are the number of pension accounts handled by them. This information may not be available by categories very readily, but the totals should not pose a problem. 20 The detailed analysis of the Civilian departments using the same methodology is available in Sen and Swain (2002).
31
Table 15: Projected Total Number of Pensioners
Years Civil Railways Telecom Post Defence Total
1999-2000 582775 1018178 100000 172022 1956243 3829218
2000-01 603070 1039364 104245 174883 2017244 3938806
2001-02 620929 1060564 108468 177871 2078564 4046396
2002-03 637449 1081491 112684 180872 2140186 4152682
2003-04 652744 1101810 116883 183980 2200645 4256062
2004-05 666916 1120500 125172 187219 2259546 4359353
2005-06 680057 1137915 133826 190495 2317084 4459377
2006-07 692253 1154101 142916 193791 2373035 4556096
2007-08 703581 1169506 152539 197181 2426973 4649780
2008-09 714111 1184461 162731 200684 2478673 4740660
2009-10 723908 1198817 173512 204022 2527857 4828116 CAGR 2.20% 1.60% 5.70% 2.10% 2.60% 2.3%
Structure of Pensions and Other Retirement Liabilities
As mentioned earlier, the time path of future pension liabilities of the
Central government would be partly driven by the increase in number of pensioners and partly by the average pension payment per pensioner. As has already been estimated in the previous section, the number of pensioners is likely to grow at an average annual rate of 2.3% over the decade, which is clearly not a very alarming figure. However, a substantial portion of the increase in pension payments would be due to the increase in the general price level since a large portion of the total pension bill is indexed to price change. Thus, even in the absence of any further pay revision, the pension liability of the Central government would increase steadily at a rate which would combine the twin effects of the increase in number of pensioners and the increase in the average pension payment on account of a general process of inflation.
A convenient way of estimating the likely future path of the aggregate
pension bill would be by regressing the annual pension payments on the stock of pensioners and a price index. However, as has already been mentioned, this path is closed to us since no reliable time series on the stock of pensioners exists, either for the total or for any individual department. Moreover, the structure of pensions, which is complex enough as it is, has undergone substantial revisions from time to time, which reduces the reliability of any time-series estimate. There is thus no option but to estimate average pensions from first principles, i.e. by working it out from the prevailing pension rules. The complexity of the pension system in the Central government requires that each and every component of the retirement benefits be separately accounted for so as to ensure accuracy and
32
precision. In order to arrive at a reasonable projection of Central government’s pension bill, it will, therefore, be necessary to understand the present pension structure and other retirement benefits applied to different categories of pensioners and the extent to which each component of pensionary benefits is driven by inflation. This section discusses the pension structure as it prevails at present.
The retirement benefits provided to the employees of all the five accounting
departments of the Central government are similar in their composition. For instance, each accounting department provides for a basic pension along with dearness relief, commutation of pension, retirement gratuity, death gratuity, leave encashment, and restoration of commutation to the pensioners who survive more than 15 years after retirement. While the basic composition of the retirement benefit is similar, the average expenditure per capita under each component varies from department to department depending upon the wage structure, service condition, age-cohort of the employees and retirement pattern. For example, average basic pension would be subject to inter-departmental variations, because of difference in service structure and average age of retirement. Rules of commutation for Defence employees are different from that of the civilian employees belonging to the departments other than Defence. Hence, it has been felt necessary to estimate the future projection of pension separately for each category of pensioners in each department. The following paragraphs describe the component with formula for estimating the pension bill. Basic Pension
As per prevailing pension rules, the monthly basic pension of a normal service pensioner for full qualifying service of 33 years is 50 per cent (pro-rated as per years of services rendered) of the average basic salary drawn by the employee during last 10 months of service before retirement. A SOFP draws 30 per cent of the average basic salary as monthly basic pension. The basic pension of the FFP is 50 per cent of the average salary for the initial 7 years. Thereafter basic pension gets reduced to 30 per cent of the basic salary. All the three variants of the basic pension are linked to one common factor i.e. the average basic salary of the retiree or deceased, as the case may be. Thus, the first step in estimating the average basic pension of different category of pensioners is to estimate the average basic salary of the government employees belonging to each accounting departments.
Our study estimates the average basic salary of government employees for each department as the weighted average of the average basic salary of different groups of Central government employees as described below.
B = Σ pibi ⁄ Σ pi ` (7)
33
where : B = average basic salary
bi = average basic pay of the employees belonging to ith class of government service i = 1,2,3,4 (same as group A, B, C & D respectively) pi = number of new retirees21 belonging to ith class of Government service
The average basic salary arrived at for each department separately as per the above-mentioned formula has been used to estimate the basic pension of all pensioners of the concerned department. In this computation, the implicit assumptions are that: (a) the group-wise distribution of total pensioners (Class I, Class II, Class III and Class IV) is same as that of new retirees or existing working population as the case may be; (b) the average pay of superannuating pensioner and deceased employee are assumed to be same;22 and (c) every pensioner has 33 years of qualifying service, which makes them eligible to draw 50 per cent of the average basic salary as pension. In case of railways pensioners, basic pension has been pro-rated for 30 years of average service. Commutation of Pension
As per prevailing pension rules for Central government employees, the
service pensioners can commute 40% (45% for Defence employees) of his/her pension after retirement in pursuance of the implementation of FCPC recommendations. Commutation of pension is a lump-sum payment made to the new retirees as per the following formula 23.
C = (0.4) ∗ (0.5) B ∗12 ∗ f (8) where: C = average value of commutation per pensioner B = average monthly basic salary f = commuted value factor (determined by age at next birthday) The commuted value factor as prescribed in the commutation table would be about 10 years (15 years for Defence) for a pensioner who submits the application for commutation within the first year of his/her retirement.
For the purposes of estimation in this study, it is assumed that every retiree applies for commutation within the first year of his/her retirement. The payment on account of commutation of pension is thus estimated by multiplying the number of new retirees by the average value of commutation per pensioner. The monthly
21 Class-wise distribution of employees has been taken for the departments where class-wise distribution of retirees is not available. 22 As has been mentioned, the average age of in-service deaths in civilian departments is 54 years as compared to the age of superannuation of 60. Thus, the average pay should be lower for the former. However, the difference is not large enough to merit any adjustment. 23 For defence employees: C = (0.45) * (0.5) * B * 12 * f
34
basic pension of the SPs has accordingly been adjusted downwards after netting out the commuted pension. Dearness Relief (DR)
The DR paid to the pensioners on the basic pension is fully linked to
inflation. DR was 37% of the basic pension in the year 1999-2000 and 43% in the year 2000-01. Our estimate incorporates the actual DR payable up to the financial year 2000-01. Thereafter, an annual price increase of 6% has been applied to the Dearness Relief for our estimation purposes on the basis of the assumption of full neutralization of inflation over the basic pension as per the FCPC effect. It is useful to note here that all the exercises undertaken so far to assess the future pension burden of Central government have either ignored the impact of price change or ignored the DR formula subsequent to commutation of pension. For the service pensioners, who receive basic pension less commutation, DR is calculated over full basic pension, and not on the post-commutation value. Death-cum-Retirement Gratuity (DCRG)
At present, retirement gratuity admissible to a pensioner is 16.5 times of the last monthly emoluments drawn (basic pay + DA) at the time of retirement if the retiree has 33 years of qualifying service to his credit. In our calculation, the estimate of retirement gratuity presumes every retiree to have completed 33 years of qualifying service. In so far as the death gratuity is concerned, the pension rule has been more liberal in that the death gratuity is almost double of the retirement gratuity pro-rated with the number of years of service one has put in. An average service period of 29 year has been presumed for the deceased government employee24 belonging to departments other than defence. For defence employees, retirement gratuity and death gratuity have been adjusted downward in conformity with 20 years of average service. The average service period for the deceased government employee has been presumed to be 15 years. This makes the death gratuity to be 29 times of the last monthly emoluments drawn by the deceased. The relevant formulae, therefore, are:25
RG = (B+DA) ∗ 16.5 (9) DG = (B+DA) ∗ 29 (10) Leave Encashment
The maximum number of days of earned leave that can be accumulated and encashed on retirement is 300 days or 10 months. Our estimate assumes every pensioner to have accumulated 300 days of leave. For a deceased employee, accumulated days of leave are taken to be equivalent of 7.5 months.
24 The mean age for in-service deaths has been estimated as 54 years. 25 These formulae are for civil ian employees; with appropriate adjustments for defence.
35
This component of retirement benefit has been clubbed with the sub-head DCRG in our estimation table. Restoration of Commutation
Till April 1985, the reduction in the monthly pension on account of commutation was a lifetime commitment. As per a Supreme Court judgment, however, the commuted amount of pension was restored after 15 years period with effect from 1st April, 1985. Accordingly, restoration of commutation has been estimated separately for the service pensioners who are likely to survive after 15 years of their retirement. This component assumes importance in the context of increased life expectancy at age 60. For computation of pensioners, who survive after 15 years of their retirement, the annual attrition rate as presented in Tables 10 to 13 has been assumed. This aspect of pension entitlements has probably been completely ignored in the estimation made by the Working Group. The component-wise post-retirement benefit, as explained above, has been computed per beneficiary under each department separately and applied to the total number of pensioners and annual retirees as the case may be to arrive at the total pension bill of each department. As has already been mentioned, the year 1999-2000 has been taken as the base year for our study and the projections on the category-wise pensioners for each department for the period 2001 to 2010 made in the previous section have been used. Validation of the Methodology
Before making any projections on the likely pension bill of the government, it is necessary to validate the methodology by comparing the base year estimate with the actual pension bill of the government as per the accounts. Unfortunately, it is not possible to do so with each item of pension entitlements, since data are not maintained in this manner for all departments. All that is available are the broad aggregates. With the assumptions and the pension structure as explained above, an estimation has been made, in the first instance, of the total pension bill of each accounting departments of the Central government for the base year 1999-2000. The base year figure of the department-wise total pension bill as per our estimates are provided in Table 16 along with the actual pension bill.
Table 16: Department-wise Pension Bill of Central Government 1999-2000
(in Rupees Crore) Departments Estimate Actual Difference (%) Civil 3248 3286 - 1.16 Railways 4056 4022 + 0.85 Post 681 677 + 0.59 Telecom 455 437 + 4.12 Defence 8433 11024 - 23.50
36
As can be observed from the table, the civil pension estimated at Rs.3248
crores for the year 1999-2000 is comparable to the actual pension bill of the central government amounting to Rs. 3285 crores. The difference is marginal, of only about 1%, which could be due to either errors arising out of our approximations or from some left over arrear payment on account of delays in giving effect to the FCPC award. The base year estimates for the Department of Post and the Railways are almost identical to the actuals. For the Telecommunication department, however, our estimate is somewhat higher than the actual, by about 4%, probably because of approximation errors. For projections this is not so much of a problem, since it will tend to overstate the likely pension bill and not to underestimate it.
As regards the Defence department, however, a substantial difference of over 23% is observed between the estimated pension bill and the actual pension bill for the year 1999-2000. A gap of this magnitude calls for further analysis since it can lead to substantial projection errors. One obvious reason could be a large component of arrear payments being included in the actual pension bill. The very fact that the Fifth Central Pay Commission award was implemented for defence only in the year 1998, and most of the benefits on account of pay revision to the pensioners were given in the year 1999-2000, there is a concentration of arrear payments in this year. In addition, the centralized system of pension administration under defence accounts, with the beneficiaries largely scattered all over the country, including in remote corners, has resulted in staggered implementation of full-scale pension revision. Hence, arrear payments in defence pension spill over to the financial years beyond 1999-2000 as well. There is evidence of significant amount of arrear bills in the years 2000-01 and 2001-02 as well.
One way to validate our estimates in the presence of arrear payments
would be to compare the actual pension bill of the defence account and our estimates continuously for a six year period starting from 1996, the year from which the pay revision became effective. Since arrear payments are supposed to more or less exactly cover the underpayments made in the earlier years, the hypothesis is that our estimates should be in excess of the actuals in the first three years and less than the actuals in the subsequent three, with the totals for the six year period being roughly equal.26 Graph 6 presented below depicts the comparative picture of the estimated pension bill and the actual accounts of defence pension from 1996-97 onwards. As can be observed, the estimated values, starting from a higher level, exhibit a smooth upward movement, whereas the actual pension payments start from a lower level and increase steeply, with a sudden surge during the years 1998 to 2000. Thus, the pattern of differences between the two series confirms our a-priori expectation that the divergence
26 It should be noted that our estimated series is based on the average pension payments calculated on the basis of the post-Pay Commission pension structure and applied to the estimated number of pensioners in each year.
37
between the estimated and the actual pension payment in defence arises essentially from arrear payments.
Graph 6: Defence Pension Bill (Rs. Crore)
0
2000
4000
6000
8000
10000
12000
1996-97 1997-98 1998-99 1999-2000* 2000-01 2001-02
Estimates
Actuals
However, one further step is necessary to validate our contention – that the excess of the estimates over the actual prior to 1998 are almost equally compensated for by the excess of actual over estimates in the three years succeeding the year 1998-99. In absolute terms, as shown in Table 17, the total pension bill as per the defence pension account (Rs 47,950 crore) for all these six years almost matches the total of the estimated pension liability (Rs 47,928 crore) for the same years, with a negligible deviation of 0.1 percent. The downward movement of the actual pension bill after 1999-00 is also indicative of the fact that defence pension is still having sizable arrear bill and there would be convergence of the actual account with the estimated number in the subsequent year itself.27
Table 17: Pension Payment to Defence Employees (Rs. Crore)
Years Estimated Actual 1996-97 6155 3683 1997-98 6801 4947 1998-99 7901 7270 1999-00 8433 11024 2000-01 9043 10539 2001-02 9595 10487
Total 47928 47950
27 In fact the RE-2002-03 and BE-2003-04 of the defence pension are Rs 10,092 crore and Rs. 11,000 crore respectively as against our estimates of Rs. 10,059 crore and Rs. 10,837 crore.
38
Projection of Pension Bills
After establishing the validity of the methodology as described above, the
pension liability of all the five accounting departments of the central government has been projected up to the year 2010 under a scenario of 6% annual inflation. These projections adopt the projected time profile of the number of pensioners in each category for all the departments computed by us as presented in the earlier section. The additional assumptions undertaken for these projections are: § The existing pension structure would prevail throughout, including full
neutralization of inflation over basic pension. § No other Pay Commission recommendation would be implemented for the
pensioners. § No presently non-pensionable categories of government employees would
be brought under pension cover.
Table 18 indicates, in summary form, the department-wise projections of the pension liability of the Central government over the ten-year period ending 2009-10. In conformity with our a priori expectations that the pension bill should grow at a rate only somewhat lower than the sum of the rate of growth of pensioners and the rate of inflation, this exercise projects the total government liabilities on account of pension payment and provision of other retirement benefit to increase at an annual average growth rate of 8.1%, with the number of pensioners growing at 2.3% per annum and an assumed inflation rate of 6%. A comparison of Table-18 with Table-15, however, shows that the postulated relationship holds for the Civil and Defence departments, but not for the departmental enterprises of Railways, Telecom and Posts, where the pension bills grow faster than the sum of the two components.
Table 18: Projected Pension Payments to Central Government Employees (Rs. Crores)
Years Civil Railways Telecom Post Defence Total 1999-2000 3250 4056 455 681 8433 16875
2000-01 3488 4669 618 866 9043 18684 2001-02 3767 5023 675 935 9595 19995 2002-03 4056 5400 736 1008 10147 21347 2003-04 4355 5789 801 1086 10994 23026 2004-05 4684 6207 1059 1171 11898 25020 2005-06 5041 6662 1186 1264 12886 27038 2006-07 5409 7128 1329 1363 13941 29170 2007-08 5833 7651 1494 1476 15070 31524 2008-09 6249 8181 1673 1592 16227 33923 2009-10 6734 8800 1884 1727 17561 36706 CAGR 7.6% 8.1% 15.3% 9.8% 7.6% 8.1%
39
The reason for this lies in the composition of the pensioners and the rate of
accretion to each of the pensioner categories. Since the average pension varies from category to category, the aggregate pension payments reflect the composition. A clearer picture emerges from the detailed break-down of the departmental pension bills. The component-wise projection of the pension bill for each accounting department is provided in the Annexure. As has been explained in the previous section, a substantial portion of the retirement benefits is indexed to inflation through payment of Dearness Relief (DR), which is applied to the basic pension. The share of basic pension-cum-dearness relief in the total pension bill for the year 1999-2000 ranges between 61 to 63 percent for all departments. The share increases to a level ranging from 69% in the Department of Post to 75% in case of defence pensions. The increased share of this component is in conformity with the assumption of full neutralization of inflation on the basic pension. In fact, this component would increase at a higher rate than the rate of inflation due to the very fact that the number of pensioners who receive the basic pension and dearness relief, would also increase annually at a rate of more than 2%. In addition, basic pension is paid to the pensioners after adjusting for the value of commutation, whereas the dearness relief is calculated on the full amount of basic pension.
Besides basic pension-cum-dearness relief, the other components of
retirement benefits such as DCRG and leave encashment are linked to inflation since these benefits are based on the last monthly emolument (basic + DA) drawn by the retiree/deceased. The gratuity is applied to the annual retirees/deceased. Hence it is possible that this component would increase at a higher rate than the annual inflation if the number of annual retirees increases over the ten-year period for which projections have been made. The only un-indexed components of the pension liability of government are commutation of pension and restored commutation. Any annual increase in these components would be attributed solely to the increased number of annual retirees. Further, government’s liability under restored commutation has a potential to increase in future because of the gradual increase in the life expectancy of the population at age 60 and above. In brief, the projected growth rate of pension liabilities of 8.1% per annum in nominal terms is consistent with the assumption of 6% annual inflation and the projected pensioners’ growth of about 2.3%. This projection is only 0.2 percentage point less than the sum of inflation rate and growth rate of pensioners. This gap is explained by the un-indexed component of pension bill, which pulled down the projection a little. Thus, the implicit elasticity of the pension bill to the inflation rate is 0.96, which is a clear indication of the extent to which pensions have become indexed post FCPC. In stark contrast, the estimate of pension bill contained in the two high level Committee reports mentioned earlier in this paper projected an annual increase of pension bill of about 4% and 4.8% during the same period, with inflation assumed to be 6% and the annual increase in the
40
number of pensioners projected to be much more than 2.3%28. The implicit elasticity as per the Working Group estimates works out to a mere 0.16, which implies a very small degree of indexation. Even a cursory perusal of the FCPC recommendations will reveal the unrealistic nature of this level of indexation.
In absolute terms, the projection of lower growth rates of the pension bill by the two high level committees cumulates to sizeable underestimation over time as compared to our estimates. In 2009-10, for instance, the Working Group has estimated a pension bill of Rs. 29,890 crore as compared to our estimate of Rs. 36,706 crore – i.e. an underestimation of 23%. The extent of underestimation in the Expert Group report is lower at about 11% but still sizable. It needs to be emphasised that inaccuracy in estimation of the pension bill, especially with these levels of underestimation, if accepted as the basis for action may lead to gross errors in budgeting and policy intervention. Conclusion
One of the more inexplicable recent decisions of the government has been the concerted move to shift away from the existing defined-benefit pension system to a defined-contribution system for all new government employees. This is especially so since both the expert bodies set up by the government to examine the likely future time path of pension liabilities have come up with projections which indicate that the pension bill is likely to grow significantly slower than even the inflation rate, let alone the expected growth of nominal GDP. Thus, these projections suggest that the pension bill as a percentage of GDP is likely to decline sharply in the coming years, which is a potent argument for not tampering with the present system. Clearly then, the government has not placed much credence on these projections while taking its decision. At the root of this obvious disbelief is probably the high growth rate of pension payments in recent years, which does not appear to be consistent with the projections. The projections made in this paper are considerably higher than the earlier estimates and suggests that the government was probably right in not entirely trusting the expert groups’ projections.
Nevertheless, our projections too do not support the need for an immediate
change in the pension system. The projected growth in the pension bill estimated at 8.1% per annum in nominal terms, as computed in our study, implies little more than 2 per cent growth in real terms. This projection should not cause much concern for an economy which has been growing at a real rate or more than 5.5% annually for last two decades. Thus, it should be possible for the economy to support the present Central government pension system under an economic scenario in which the economy maintains the historical growth path. The position could be even more comfortable in the future if the Indian economy grows at an 28 The Working Group Report (2001) projects an annual growth of more than 3 per cent in number of pensioners. The Expert Group Report (2002) does not provide the projected pensioners’ number.
41
accelerated rate, as is being widely anticipated. Thus, there does not appear to be any imminent danger of the pension bill going out of control even under the present pension system.
There is, however, an important caveat that needs to be noted, since it has
the potential for seriously disturbing the projections. The net addition to the stock of pensioners is the outcome of accretions arising out of retirement and attrition due to deaths.29 The composition of this process of change is illustrated for the case of service pensioners (SPs) in Table 19. As can be seen, the rate of net addition is significantly lower than the rate of annual accretion in all cases due to fairly high attrition rates. If, for whatever reason, pension accounts are not closed or converted on the death of the pensioner, the net addition figure could rise significantly and, at the limit, may even approach the gross accretion rates. In such a situation, the rate of growth of the pension bill could possibly even exceed the rate of growth of nominal GDP.
Table 19: Composition of Change in Service Pensioners
(average annual change 1999-00 to 2009-10) Civil Railways Telecom Post Defence Total
Accretion 4.9% 6.7% 11.5% 7.0% 4.0% 5.2% Attrition 3.2% 5.5% 4.8% 5.0% 1.5% 3.0% Net Addition 1.7% 1.2% 6.7% 2.0% 2.5% 2.2%
As far as gross accretions are concerned, the systems which are in place
for issuance of pension payment orders appear adequate for ensuring that no major misuse takes place.30 The fairly close correspondence between the model-based estimates of retirees and the actual PPOs issued is a confirmation of this judgement. On the other hand, the system for closure of pension accounts is weak at best, and there is ample scope for manipulation. Thus, the possibility certainly exists that the actual rates of attrition may fall well below estimated, which would lead to higher growth of pension payments than projected in this paper.
In conclusion, therefore, the macroeconomic sustainability of the Central
government pension bill under the current pension system will depend on the following policy and administrative considerations being met:
§ Employment pattern of government employees would remain stable
with no future growth. In fact, all the five Central government departments discussed in our study seem to have started
29 Equations (4), (5) and (6) clearly bring out the nature of this process, whereby for each category of pensioners there is a well-defined source of annual accretion and attrition. 30 Although, from time to time, cases of fraudulent PPOs are uncovered, the numbers do not appear to be large.
42
consolidating their position on employees’ number, which has either remained constant or declined marginally.
§ No upward revision in the pension structure or in the coverage of
government employees would take place in the future.
§ Pension administration would be efficient enough to maintain appropriate lists of all categories of pensioners and timely deletion of the names in the event of their attrition from the pensioners’ list. This is one area which is of serious concern as of now and requires significant improvement.
Thus, in the absence of further upward pay revision or in the pension structure and with improved system of pension administration, the pension bill of the Central government should be economically sustainable. The present concern relating to the government pension system is perceived more in terms of the governments’ ability to sustain the Pay As You Go (PAYG) system of pension, in which the entire pension bill is charged to the Consolidated Fund of India and met out of the revenue receipt of the Central government. Our assessment of the future pension bill indicates that if the buoyancy of Central government revenue receipt can be maintained at least at 1.00, then the claim of the pension bill on revenue receipt would decline steadily. The fiscal sustainability argument in favour of the on-going pension reform, therefore, seems to be rather weak.
43
ANNEXURE
Projection of Pension Payments at 6% Annual Rate of Inflation
Railways
(Rs. Crores) Years Basic Pension+DR Commutation Retiremt Death Restored PENSION
of Gratuity Gratuity Commutn BILL Service SOFP FFP Pension (Service) (FFP) Total 1999-2000 1753 615 343 411 593 128 212 4056 2000-01 1889 671 360 562 846 134 208 4669 2001-02 2085 740 383 565 902 142 206 5023 2002-03 2282 818 407 571 965 151 206 5400 2003-04 2485 901 432 574 1029 160 208 5789 2004-05 2711 986 460 576 1093 169 213 6207 2005-06 2959 1075 488 578 1164 179 218 6662 2006-07 3210 1169 519 579 1235 190 226 7128 2007-08 3501 1271 551 580 1312 202 234 7651 2008-09 3775 1381 585 583 1398 214 244 8181 2009-10 4111 1501 621 588 1495 226 257 8800 CAGR 8.9% 9.3% 6.1% 3.6% 9.7% 5.8% 2.0% 8.1%
Defence
(Rs. Crore) Years Basic pension+DR Commutation Retiremt Death Restored PENSION
of Gratuity Gratuity Commutn BILL Service SOFP FFP Pension (Service) (FFP) Total
1999-2000 3917 608 806 1249 803 191 859 8433 2000-01 4338 657 869 1252 840 199 889 9043 2001-02 4682 712 933 1260 881 207 921 9595 2002-03 5038 765 993 1267 916 214 954 10147 2003-04 5604 841 1083 1275 976 227 987 10994 2004-05 6218 920 1180 1281 1040 241 1018 11898 2005-06 6882 1004 1283 1287 1108 255 1066 12886 2006-07 7601 1091 1393 1291 1178 271 1116 13941 2007-08 8381 1181 1507 1295 1252 287 1168 15070 2008-09 9224 1272 1628 1298 1330 256 1221 16227 2009-10 10136 1364 1752 1300 1412 322 1274 17561 CAGR 10.0% 8.4% 8.1% 0.4% 5.8% 5.4% 4.0% 7.6%
44
POST
(Rs. Crores) Basic pension+DR Commutation Retiremt Death Restored PENSION
Years of Gratuity Gratuity Commutn BILL Service SOFP FFP Pension (Service) (FFP) Total
1999-2000 341 121 41 41 61 24 51 681 2000-01 368 129 43 97 154 26 50 866
2001-02 407 139 46 100 167 27 49 935 2002-03 447 150 50 102 182 29 48 1008 2003-04 490 161 53 105 198 30 48 1086 2004-05 538 173 57 107 214 48 1171 2005-06 593 185 61 110 232 34 49 1264 2006-07 650 198 65 112 251 36 50 1363 2007-08 718 212 69 115 272 38 52 1476 2008-09 785 226 74 117 295 41 54 1592 2009-10 868 240 78 121 322 43 56 1727 CAGR 9.8% 7.1% 6.7% 11.4% 18.0% 5.8% 0.9% 9.8%
Telecommunications
(Rs.Crore) Years Basic pension+DR Commutation Retirement Death Restored PENSION
of Gratuity Gratuity Commutn BILL Service SOFP FFP Pension (Service) (FFP) Total
1999-2000 204 47 63 34 51 38 17 455
2000-01 225 63 68 79 125 40 17 618
2001-02 254 71 74 81 136 42 17 675
2002-03 284 78 81 84 149 43 17 736
2003-04 316 86 89 86 162 45 18 801
2004-05 372 95 96 144 287 47 18 1059
2005-06 437 104 105 152 321 48 20 1186
2006-07 511 114 113 161 360 50 21 1329
2007-08 600 124 122 170 403 52 24 1494
2008-09 696 137 131 179 451 54 26 1673
2009-10 816 149 141 189 504 55 30 1884 CAGR 14.8% 12.2% 8.4% 18.7% 25.7% 3.8% 5.8% 15.3%
45
Civil (Rs. Crore)
Years Basic pension+DR Commutation Retirement Death Restored PENSION of Pension Gratuity Gratuity Commutn BILL Service SOFP FFP (Service) (FFP) Total 1999-2000* 1546 248 225 312 472 174 275 3250 2000-01 1677 267 242 346 546 129 283 3488 2001-02 1861 291 264 346 578 137 290 3767 2002-03 2049 318 287 346 145 298 4056 2003-04 2242 346 311 346 650 153 306 4355 2004-05 2457 378 339 346 689 163 313 4684 2005-06 2695 412 366 346 730 172 320 5041 2006-07 2936 449 393 346 774 183 328 5409 2007-08 3219 489 426 346 821 194 339 5833 2008-09 3489 533 461 346 870 205 346 6249 2009-10 3818 580 499 346 922 218 352 6734 CAGR 9.5% 8.9% 8.3% 1.0% 6.9% 2.3% 2.5% 7.6%
46
References
Government of India (2001), “An Assessment of Government of India’s Pensionary Liability”, Report of the Working Group, Ministry of Finance, New Delhi. Government of India (2002), “Report of High Level Expert Group on New Pension System”, Ministry of Finance, New Delhi. Sen, Pronab & Swain, Sibani (2002), “Technical Study on Retirement and Pensions: Projections of The Central Government”, Working Paper No. 1/2002-PC, Planning Commission, Government of India.
